LODES Data - Tyler Yankee

Load packages

library(tidyverse)
library(lehdr)

Warning: package ‘lehdr’ was built under R version 4.2.2

library(sf)
library(tmap)
library(tidycensus)
library(janitor)
library(here)
library(lubridate)

Load LODES Data

See the 2019 version of the technical document for descriptions of what the variables mean and other information: https://lehd.ces.census.gov/data/lodes/LODES7/LODESTechDoc7.4.pdf.

Three kinds of data:

• “od” - origin to destination (info on both residence and workplace of workers)
• “rac” - residential information of workers
• “wac” - workplace information of workers

Some quick examples of data codes for od:

• w/r_geocode: workplace/residential census block codes
• S000: total number of jobs
• SA01-03: number of jobs broken down by age groups
• SE01-03: number of jobs broken down by monthly income groups
• SI01-03: number of jobs broken down by industry groups

dc_od <- grab_lodes(state ="dc", year = 2019, lodes_type = "od", job_type = "JT01", 
           segment = "S000", state_part = "aux", agg_geo = "tract",use_cache=TRUE)

Cached version of file found in C:\Users\tsyan\AppData\Local/lehdr/lehdr/Cache/dc_od_aux_JT01_2019.csv.gz
Reading now...

dc_wac <- grab_lodes(state = c("dc"), year = 2019, lodes_type = "wac", job_type = "JT00", 
           segment = "S000", state_part = "aux", agg_geo = "tract") %>% filter(state=="DC")

Downloading https://lehd.ces.census.gov/data/lodes/LODES7/dc/wac/dc_wac_S000_JT00_2019.csv.gz to C:\Users\tsyan\AppData\Local/lehdr/lehdr/Cache/dc_wac_S000_JT00_2019.csv.gz now...
Removed C:\Users\tsyan\AppData\Local/lehdr/lehdr/Cache/dc_wac_S000_JT00_2019.csv.gz.

dc_rac <- grab_lodes(state = c("dc"), year = 2019, lodes_type = "rac", job_type = "JT00", 
           segment = "S000", state_part = "aux", agg_geo = "tract") %>% filter(state=="DC")

Downloading https://lehd.ces.census.gov/data/lodes/LODES7/dc/rac/dc_rac_S000_JT00_2019.csv.gz to C:\Users\tsyan\AppData\Local/lehdr/lehdr/Cache/dc_rac_S000_JT00_2019.csv.gz now...
Removed C:\Users\tsyan\AppData\Local/lehdr/lehdr/Cache/dc_rac_S000_JT00_2019.csv.gz.

How many people work in DC? Note the difference between the two sums — # or workers vs # of workers who live in that area.

sum(dc_wac$C000)

[1] 688073

sum(dc_rac$C000)

[1] 319855

Load census data

#census_api_key("4c7e5b613afdf7d2fd6d80005f53282c9a7c1bae")
v20 = load_variables(2019,"acs5",cache=TRUE)
df_census=get_acs(geography = "tract",
                  variables=c("total_pop"="B01003_001",
                              "worker_living"="B08301_001"),
                  state="DC",geometry=TRUE,year=2019)

Getting data from the 2015-2019 5-year ACS
Downloading feature geometry from the Census website.  To cache shapefiles for use in future sessions, set `options(tigris_use_cache = TRUE)`.
Using FIPS code '11' for state 'DC'

  |                                                                                      
  |                                                                                |   0%
  |                                                                                      
  |=================================                                               |  41%
  |                                                                                      
  |================================================================================| 100%

df_cens=df_census %>% rename(val=estimate) %>%  
  pivot_wider(names_from = "variable",
              values_from = c("val","moe")) %>%
  clean_names()

Merge census and lodes data

By joining wac data to census data, we now have a dataframe with all the data AND the geometry by tract.

df_c = df_cens %>% left_join(dc_wac, by = c("geoid" = "w_tract"))

Skufca sample visualizations

Where do men and women work?

tmap_mode("view")

tmap mode set to interactive viewing

df_c1 = df_c %>% select(pop = val_total_pop, men = CS01, women = CS02) %>%
  mutate(ratio=women/men)
df_c1 %>% tm_shape() + 
  tm_polygons(c("pop", "men", "women"), alpha = .4)

Examine the ratio between the men and women groups rather than visualizing them separately; this is a more useful result on which to base conclusions.

df_c1 %>% tm_shape() + tm_polygons(c("ratio"), alpha = .4)

Look at racial/ethnic differences in where people work.

df_c2 = df_c %>% select(white_workers = CR01, black_workers = CR02, 
                        hispanic_workers = CT02)
df_c2 %>% tm_shape() + 
  tm_polygons(c("white_workers", "black_workers", "hispanic_workers"), alpha = .4)

My visualizations

Analyze age, income, and education level distributions.

Starting with age:

df_c3 = df_c %>% select(a29 = CA01, a3054 = CA02, a55 = CA03)

df_c3 %>% tm_shape() + tm_polygons(c("a29", "a3054", "a55"), alpha = .4)

…then income:

df_c4 = df_c %>% select(i1250 = CE01, i12513333 = CE02, i3334 = CE03)
df_c4 %>% tm_shape() + tm_polygons(c("i1250", "i12513333", "i3334"), alpha = .4)

…then education level.

df_c5 = df_c %>% select(elhs = CD01, ehs = CD02, eassoc = CD03, ebac = CD04)
df_c5 %>% tm_shape() + tm_polygons(c("elhs", "ehs", "eassoc", "ebac"), alpha = .4)